Figure 134

Overall quality of essay writing and cognitive effort as a function of topic knowledge (low vs. high). Data from Kellogg (1987).

2. Sociocultural knowledge: information about the social background or context.

3. Metacognitive knowledge: knowledge about what one knows.

There is generally little relationship between the writer's knowledge of the topic (mainly conceptual knowledge) and the quality of the text produced. For example, Kellogg (1987) asked his participants to write a persuasive essay on why the United Nations should remain in New York. Those with much relevant knowledge did not write better essays than those with less knowledge. However, Kellogg (1987) also obtained a measure of cognitive effort by measuring time to respond to an occasional stimulus during the writing process. Participants with more knowledge devoted less effort to writing (see Figure 13.4) According to Hayes and Flower (1986), strategic knowledge also plays a major role in the construction of a writing plan. Strategic knowledge concerns ways of organising the goals and sub-goals of writing to construct a coherent writing plan. Hayes and Flower found that good writers use strategic knowledge very flexibly. The structure of the writing plan often changes as dissatisfaction grows with the original plan.

Adults possessing either a lot of knowledge or relatively little on a topic were compared by Hayes and Flower (1986). The experts produced more goals and sub-goals, and so constructed a more complex overall writing plan. In addition, the various goals of the experts were much more interconnected.

Planning often takes the form of writing notes. Kellogg (1994) identified three types of preliminary note-writing:

1. Clustering or networking ideas and their interrelationships, sometimes referred to as "mind maps".

2. Listing ideas, with some attempt to place them in the appropriate order.

3. Outlining ideas and how they are related hierarchically.

Sentence generation

The gap between the writing plan and the actual writing of sentences is usually large. Kaufer, Hayes, and Flower (1986) found that essays were always at least eight times longer than outlines. In some ways, the process going on here is the opposite of that involved in comprehension. Kintsch and van Dijk (1978)

argued that comprehension involves extracting the macro-structure from the micro-propositions (see Chapter 12). In contrast, sentence generation can be regarded as generating micro-propositions from the macro-structural plans.

The technique of asking writers to think aloud permitted Kaufer et al. (1986) to explore the process of sentence generation. Here is a typical verbal protocol of a writer engaged in writing, with dots indicating pauses of at least two seconds:

The best thing about it is (1) what? (2) Something about using my mind (3)

it allows me the opportunity to (4) uh I want to write something about my ideas (5) to put ideas into action (6) or to develop my ideas into (7) what? (8) into a meaningful form? (9) Oh, Bleh! say it allows me (10) to use (11) Na allows me scratch that. The best thing about it is that it allows me to use (12) my mind and ideas in a productive way (13).

In this protocol, fragments 12 and 13 formed the written sentence, and the earlier fragments 1, 4, 6, 7, 9, and 11 were attempts to produce parts of the sentence.

Kaufer et al. (1986) compared the sentence-generation styles of expert and average writers. Both groups accepted about 75% of the sentence parts they verbalised. The length of the average sentence part was 11.2 words for the expert writers compared with 7.3 words for the average writers. Thus, good writers make use of larger units or "building blocks" than do others.


According to Hayes and Flower (1986, p. 110), "the more expert the writer, the greater the proportion of writing time the writer will spend in revision." Why is this? Expert writers focus on the coherence and structure of the arguments expressed, whereas non-expert writers focus on individual words and phrases. It is much more time-consuming to modify the hierarchical structure of a text than to change individual words.

Faigley and Witte (1983) compared the revisions made by writers at different levels of skill. They discovered that 34% of the revisions by experienced adult writers involved a change of meaning, against only 12% of the revisions of inexperienced college writers. This difference probably occurred because experienced writers are more concerned with coherence and meaning.

Further evidence on differences between expert and non-expert writers was obtained by Hayes et al. (1985). Expert writers detected 60% more problems in a text than did non-experts. The expert writers correctly identified the nature of the problem in 74% of cases, against only 42% for the non-expert writers.

One of the greatest problems is make a text more comprehensible to the intended readers. This is a real problem in writing a textbook such as this, where the readers vary considerably in their previous knowledge of the topics being discussed. An interesting way of teaching writers to be more alert to the reader's needs was used by Schriver (1984). Students read an imperfect text, and predicted the comprehension problems a reader would have with it. Then the students read a reader's verbal protocol produced while he or she tried to understand that text. After the students had been given various texts plus reader's protocols, they became better at predicting the kinds of problems readers would have with new texts.


Hayes and Flower (1980, 1986) have enhanced our understanding of writing processes. However, protocol analysis (on which Hayes and Flower rely heavily) can provide information only about those processes of which there is conscious awareness. Writers are unlikely to be aware of how they search long-term memory for ideas, how they think of inferences, and so on. The requirement to verbalise while writing adds to the writer's processing load, and so may alter the writing process. Rymer (1988) found that only five out of nine scientists approached were willing to try thinking out loud while composing a scientific paper, and only one of them produced useful protocols. Directed retrospection provides less information than protocol analysis, and shares with it a focus on conscious processes. However, it is much less intrusive.

The comparison of writers having more and less skill by Hayes and Flower facilitates the identification of the specific strategies involved in skilled writing. It is also useful in terms of producing practical advice for those who find it hard to develop adequate writing skills.

On the negative side, the three processes of planning, sentence generation, and revision cannot be neatly separated. In particular, planning and sentence generation are often almost inextricably bound up with each other. A further criticism was raised by Kellogg (1990, p. 376), who argued that writing is more of a social act than was proposed by Hayes and Flower (1986).

Writing expertise

Why are some writers more skilful than others? Individual differences in writing ability depend most on the processes involved in the planning stage. Bereiter and Scardamalia (1987) identified two major strategies used in the planning stage:

• A knowledge-telling strategy.

• A knowledge-transforming strategy.

The knowledge-telling strategy involves writers simply writing down everything they know about a topic with no planning. The text already generated provides retrieval cues for generating the rest of the text. In the words of a 12-year-old child who used the knowledge-telling strategy (Bereiter & Scardamalia, 1987, p. 9), "I have a whole bunch of ideas and write them down until my supply of ideas is exhausted. Then I might try to think of more ideas up to the point when you can't get any more ideas that are worth putting down on paper."

The knowledge-transforming strategy involves use of a rhetorical problem space and a content problem space. Rhetorical problems relate to the achievement of the goals of the writing task (e.g., "Can I strengthen the argument?"), and content problems relate to the specific information to be written down (e.g., "The case of Smith vs. Jones strengthens the argument"). There should be movement of information in both directions between the content space and the rhetorical space. However, this happens more with skilled writers using a knowledge-transforming strategy. According to Bereiter and Scardamalia (1987, p. 303):

The key requirement for reflective thought in the translation of problems encountered in the rhetorical space back into subgoals to be achieved in the content space. the novice possesses productions for transferring information from the content space to the rhetorical space, but lacks productions for the return trip.

Bereiter, Burtis, and Scardamalia (1988) hypothesised that knowledge-transforming strategists would be more likely than knowledge-telling strategists to produce high-level main points capturing important

Percentage of writing time devoted to planning (P), sentence generation (SG), and revision (R) as a function of whether an outline had been produced. Based on data in Kellogg (1988).

Stop Anxiety Attacks

Stop Anxiety Attacks

Here's How You Could End Anxiety and Panic Attacks For Good Prevent Anxiety in Your Golden Years Without Harmful Prescription Drugs. If You Give Me 15 minutes, I Will Show You a Breakthrough That Will Change The Way You Think About Anxiety and Panic Attacks Forever! If you are still suffering because your doctor can't help you, here's some great news...!

Get My Free Ebook

Post a comment